You’re out of free articles.
Log in
To continue reading, log in to your account.
Create a Free Account
To unlock more free articles, please create a free account.
Sign In or Create an Account.
By continuing, you agree to the Terms of Service and acknowledge our Privacy Policy
Welcome to Heatmap
Thank you for registering with Heatmap. Climate change is one of the greatest challenges of our lives, a force reshaping our economy, our politics, and our culture. We hope to be your trusted, friendly, and insightful guide to that transformation. Please enjoy your free articles. You can check your profile here .
subscribe to get Unlimited access
Offer for a Heatmap News Unlimited Access subscription; please note that your subscription will renew automatically unless you cancel prior to renewal. Cancellation takes effect at the end of your current billing period. We will let you know in advance of any price changes. Taxes may apply. Offer terms are subject to change.
Subscribe to get unlimited Access
Hey, you are out of free articles but you are only a few clicks away from full access. Subscribe below and take advantage of our introductory offer.
subscribe to get Unlimited access
Offer for a Heatmap News Unlimited Access subscription; please note that your subscription will renew automatically unless you cancel prior to renewal. Cancellation takes effect at the end of your current billing period. We will let you know in advance of any price changes. Taxes may apply. Offer terms are subject to change.
Create Your Account
Please Enter Your Password
Forgot your password?
Please enter the email address you use for your account so we can send you a link to reset your password:
AI has already changed weather forecasting forever.
It’s been a wild few years in the typically tedious world of weather predictions. For decades, forecasts have been improving at a slow and steady pace — the standard metric is that every decade of development leads to a one-day improvement in lead time. So today, our four-day forecasts are about as accurate as a one-day forecast was 30 years ago. Whoop-de-do.
Now thanks to advances in (you guessed it) artificial intelligence, things are moving much more rapidly. AI-based weather models from tech giants such as Google DeepMind, Huawei, and Nvidia are now consistently beating the standard physics-based models for the first time. And it’s not just the big names getting into the game — earlier this year, the 27-person team at Palo Alto-based startup Windborne one-upped DeepMind to become the world’s most accurate weather forecaster.
“What we’ve seen for some metrics is just the deployment of an AI-based emulator can gain us a day in lead time relative to traditional models,” Daryl Kleist, who works on weather model development at the National Oceanic and Atmospheric Administration, told me. That is, today’s two-day forecast could be as accurate as last year’s one-day forecast.
All weather models start by taking in data about current weather conditions. But from there, how they make predictions varies wildly. Traditional weather models like the ones NOAA and the European Centre for Medium-Range Weather Forecasts use rely on complex atmospheric equations based on the laws of physics to predict future weather patterns. AI models, on the other hand, are trained on decades of prior weather data, using the past to predict what will come next.
Kleist told me he certainly saw AI-based weather forecasting coming, but the speed at which it’s arriving and the degree to which these models are improving has been head-spinning. “There's papers coming out in preprints almost on a bi-weekly basis. And the amount of skill they've been able to gain by fine tuning these things and taking it a step further has been shocking, frankly,” he told me.
So what changed? As the world has seen with the advent of large language models like ChatGPT, AI architecture has gotten much more powerful, period. The weather models themselves are also in a cycle of continuous improvement — as more open source weather data becomes available, models can be retrained. Plus, the cost of computing power has come way down, making it possible for a small company like Windborne to train its industry-leading model.
Founded by a team of Stanford students and graduates in 2019, Windborne used off-the-shelf Nvidia gaming GPUs to train its AI model, called WeatherMesh — something the company’s CEO and co-founder, John Dean, told me wouldn’t have been possible five years ago. The company also operates its own fleet of advanced weather balloons, which gather data from traditionally difficult-to-access areas.
Standard weather balloons without onboard navigation typically ascend too high, overinflate, and pop within a matter of hours (thus becoming environmental waste, sad!). Since it’s expensive to do launches at sea or in areas without much infrastructure, there’s vast expanses of the globe where most balloons aren’t gathering any data at all.
Satellites can help, of course. But because they’re so far away, they can’t provide the same degree of fidelity. With modern electronics, though, Windborne found it could create a balloon that autonomously changes altitude and navigates to its intended target by venting gas to descend and dropping ballast to ascend.
“We basically took a lot of the innovations that lead to smartphones, global satellite communications, all of the last 20 years of progress in consumer electronics and other things and applied that to balloons,” Dean told me. In the past, the electronics needed to control Windborne’s system would have been too heavy — the balloon wouldn’t have gotten off the ground. But with today’s tiny tech, they can stay aloft for up to 40 days. Eventually, the company aims to recover and reuse at least 80% of its balloons.
The longer airtime allows Windborne to do more with less. While globally there are more than 1,000 conventional weather balloons launched every day, Dean told me, “We collect roughly on the order of 10% or 20% of the data that NOAA collects every day with only 100 launches per month.” In fact, NOAA is a customer of the startup — Windborne already makes millions in revenue selling its weather balloon data to various government agencies.
Now, with a potentially historic hurricane season ramping up, Windborne has the potential to provide the most accurate data on when and where a storm will touch down.
Earlier this year, the company used WeatherMesh to run a case study on Hurricane Ian, the Category 5 storm that hit Florida in September 2022, leading to over 150 fatalities and $112 billion in damages. Using only weather data that was publicly available at the time, the company looked at how accurately its model (had it existed back then) would have tracked the hurricane.
Very accurately, it turns out. Windborne’s predictions aligned neatly with the storm’s actual path, while the National Weather Service’s model was off by hundreds of kilometers. That impressed Khosla Ventures, which led the company’s $15 million Series A funding round earlier this month. “We haven’t seen meaningful innovation in weather since The Weather Channel in the 90s. Yet it’s a $100 billion market that touches essentially every industry,” Sven Strohband, a partner and managing director at Khosla Ventures, told me via email.
With this new funding, Windborne is scaling up its fleet of balloons as it prepares to commercialize. The money will also help Windborne advance its forecasting model, though Dean told me robust data collection is ultimately what will set the company apart. “In any kind of AI industry, whoever has the top benchmark at any given time, it’s going to fluctuate,” Dean said. “What matters is the model plus the unique datasets.”
Unlike Windborne, the tech giants with AI-based weather models — including, most recently, Microsoft — aren’t gathering their own data, instead drawing solely on publicly accessible information from legacy weather agencies.
But these agencies are starting to get into the game, too. The European Centre for Medium-Range Weather Forecasts has already created its own AI-based model, the Artificial Intelligence/Integrated Forecasting System, which it runs in parallel to its traditional model. NOAA, while a bit behind, is also looking to follow suit.
“In the end, we know we can't rely on these big tech companies to just keep developing stuff in good faith to give to us for free,” Kleist told me. Right now, many of the top AI-based weather models are open source. But who knows if that will last? “It's our mission to save lives and property. And we have to figure out how to do some of this development and operationalize it from our side, ourselves,” Kleist said, explaining that NOAA is currently prototyping some of its own AI-based models.
All of these agencies are in the early stages of AI modeling, which is why you likely haven’t noticed weather predictions making a pronounced leap in accuracy as of late. It’s all still considered quite experimental. “Physical models, the pro is we know the underlying assumptions we make. We understand them. We have decades of history of developing them and using them in operational settings,” Kleist told me. AI-based models are much more of a black box, and there’s questions surrounding how well they will perform when it comes to predicting rare weather events, for which there might be little to no historical data for the model to reference.
That hesitation might not last long, though. “To me it’s fairly obvious that most of the forecasts that would actually be used by users in the future will come from machine learning models,” Peter Dueben, head of Earth systems modeling at the European Centre for Medium Range Weather Forecasting, told me. “If you just want to get the weather forecast for the temperature in California tomorrow, then the machine learning model is typically the better choice,” he added.
That increased accuracy is going to matter a lot, not just for the average weather watcher, but also for specific industries and interest groups for whom precise predictions are paramount. “We can tailor the actual models to particular sectors, whether it's agriculture, energy, transportation,” Kleist told me, “and come up with information that's going to be at a very granular, specific level to a particular interest.” Think grid operators or renewable power generators who need to forecast demand or farmers trying to figure out the best time to irrigate their fields or harvest crops.
A major (and perhaps surprising) reason this type of customization is so easy is because once AI-based weather models are trained, they’re actually orders of magnitude cheaper and less computationally intensive to run than traditional models. All of this means, Kleist told me, that AI-based weather models are “going to be fundamentally foundational for what we do in the future, and will open up avenues to things we couldn't have imagined using our current physical-based modeling.”
Log in
To continue reading, log in to your account.
Create a Free Account
To unlock more free articles, please create a free account.
It will take years, at least, to reconstitute the federal workforce — and that’s if it can be managed at all.
By anyone’s best guess, there are — or soon will be — 284,186 fewer federal employees and contractors than there were on January 19, 2025. While Voice of America and the U.S. Agency for International Development have had it the worst, the Trump administration’s ongoing reductions have spared few government agencies. Over 10% of the staff at the National Oceanic and Atmospheric Administration, including at critical weather stations and tsunami monitoring centers, have left or been pushed out. Layoffs, buyouts, and early retirements have reduced the Department of Energy’s workforce by another 13%.
The best-case scenario for the civil service at this point would be if the administration has an abrupt change of heart and pivots from the approach of government “efficiency” guru Elon Musk and Office of Management and Budget Director Russell Vought, who has said he wants government bureaucrats to be “traumatically affected” by the funding cuts and staff reductions. Short of that unlikelihood, its membership will have to wait out the three-and-a-half remaining years of President Trump’s term in the hopes that his successor will have a kinder opinion of the federal workforce.
But even that wouldn’t mean a simple fix. In my effort to learn how long it would take the federal workforce to recover from just the four-plus months of Trump administration cuts so far, no one I spoke to seemed to believe a future president could reverse the damage in a single four-year term. “It will be very difficult, if not impossible, to restore the kind of institutional knowledge that’s being lost,” Jacqueline Simon, policy director of the American Federation of Government Employees, the largest union of federal government workers, told me.
There are three main reasons why restaffing the government will be trickier than implementing a simple policy change. The first is that the government had already been strugglingto fill empty posts before Trump’s layoffs began. “For a considerable period of time, the biggest challenge for the federal government, in personnel terms, has been getting talented people into government quickly,” Don Moynihan, a professor at the Ford School of Public Policy at the University of Michigan, told me. “That was already a problem preceding the Trump administration, and they just made it a lot worse.”
Before Trump’s second term, an estimated 83% of “major federal departments and agencies” struggled with staff shortages, while 63% reported “gaps in the knowledge and skills of their employees,” according to research by the Partnership for Public Service, a nonprofit supporting the civil service. Even President Joe Biden, who’d promised to restore a “hollowed out” federal workforce after Trump 1.0, struggled at the task, ultimately growing the number of permanent employees by just 0.9% by March 2023. (He eventually saw 6% growth over his entire term; a bright spot was hiring for roles necessary for carrying out the Infrastructure Investment and Jobs Act.)
Still, as I’ve previously reported, many hard-to-fill roles in remote locations or that required specialized skills were empty when Trump came into office and ordered a hiring freeze.
The second challenge to rebuilding the federal workforce is that many employees who have left the government may not be able to — or may not want to — return to their previous roles. Staff who have taken early retirements will be permanently lost or have to return as rehired annuitants, which Simon of the American Federation of Government Employees noted has “a lot of disadvantages,” including, in some cases, earning less than the minimum wage. Other former employees, particularly in the sciences, may have been enticed abroad as part of the U.S. brain drain. Still others may have found enjoyable and fulfilling work at the state level, in nonprofits, or in the private sector, and have no interest in returning to government.
It certainly doesn’t help that the Trump administration has made the federal government a less competitive employer. Abigail Haddad, a data scientist for the Department of the Army and, until recently, the Department of Homeland Security’s AI Corps, wrote for Moynihan’s Substack,Can We Still Govern?, that she’d been hired for a fully remote job, only to be told “we would be fired if we did not immediately return to office 9 to 5, five days a week.” Rather than make a two-and-a-half-hour round-trip commute to “an office that was never mentioned when I took the job,” Haddad quit. “It was clear to me that the people making these decisions about my work conditions were not only unconcerned about my ability to be productive, but were actively hostile toward it,” she wrote.
The last obstacle to reversing the Trump administration’s cuts echoes Haddad’s experience — and is, in my view, the most worrisome of all. That is, the current landscape will almost certainly dissuade future generations from pursuing jobs in the government. “There will be some opportunities in states and nonprofits,” Simon noted. “But as far as an opportunity for public service in the federal government — they’ve made that an impossibility, at least for the next many years.”
Moynihan, the public policy professor, added that while it’s still early to predict what students will do, he’s heard worries in his classrooms about “what future job prospects look like, given the instability around the federal government.” But the crisis goes beyond just hiring concerns.
“There’s a whole generation of public servants who would say they were inspired to go into government because they heard John F. Kennedy say, ‘Ask not what your country can do for you — ask what you can do for your country,’” he said. “There is a genuine value in elected leaders calling on people to serve and presenting that service in noble terms.” Most people don’t join the public sector for the paycheck, after all — it’s for the “opportunity to do meaningful work, and for job stability and security,” Moynihan went on. The Trump administration has gutted the promises of both.
So then, how long would it take to restaff the government? Simon told me that since it was an executive order that directed the cuts, they could be functionally undone by another executive order, though the rehiring process itself “could take years.” Moynihan used the metaphor of a muscle, rather than a switch that gets turned on and off, to answer the same question. “The Trump administration is cutting a lot of muscle right now, and so the next president will not be able to simply, on day one, bring that back,” he told me. “They’ll have to be able to persuade people that the workspace is no longer going to be toxic, is going to be more secure, and will allow them to do meaningful work — and they’re going to face a fairly skeptical audience, given everything that’s going on.”
But that’s if things hold as they are. They could still get worse.
As the administration continues its attack on the civil service, it seems all but sure to be cueing up an eventual Supreme Court case over the legality of reclassifying federal employees so that they can be easily fired if they’re perceived as not loyal enough to the president. And if the court rules that the president can do so, “any sort of law that Congress might put in the future that constrains those powers is unconstitutional,” Moynihan said. In that scenario, the government would no longer be able to provide “any sort of long-term credible commitments to potential employees that four years down the line or eight years down the line, any new president could just rip up their workplace” or lay them off for arbitrary reasons.
The answer to how long it would take to restaff the federal government after Trump, then, takes on an entirely different tenor — it may never be the same again.
Smothered, covered, and recharged.
Picture, if you will, the perfect electric vehicle charging stop. It sits right off a well-traveled highway. It has decent bathrooms, preferably ones that are open 24/7. It gives drivers and road-tripping families a simple way to occupy themselves during the 15 to 30 minutes it takes to refill the battery, the most obvious solution being a meal that can be consumed within that time window.
In other words, it is a Waffle House.
The beloved chain of budget restaurants spread across the American South said last week that it would begin to install DC fast chargers in 2026. Built by BP, the charging stalls will be able to deliver up to 400 kilowatts of electricity and will include plugs with both the Combined Charging System standard (the plug used by most non-Tesla EVs to date) and the North American Charging System standard (the formerly proprietary Tesla plug that is slowly becoming the standard for the industry at large). At last, Americans can get their hash browns smothered, covered, and recharged.
We won’t see every Waffle House in the country become an electron depot overnight. BP said it is planning installations at about 50 sites right now; Waffle House has around 2,000 locations in the United States. Yet the addition of charging — and not just charging, but high-speed charging — at the Waffle House is just what the American EV experience needs.
Where fast chargers are built has been driven by a few factors. Notably, there is necessity from the EV driver’s point of view and practicality for the charging company. Charging depots along major highways and interstates make electric road trips possible, but many prime pit stops between big cities are in the middle of nowhere, which makes it a challenge to provide amenities to resting drivers. In the empty California desert between L.A. and San Francisco, for example, Tesla built Superchargers at iconic steak restaurants and at existing travel plazas with your expected array of gas stations and fast food restaurants. I’ve also stopped numerous times at an impromptu, formerly unpaved site rushed together to accommodate holiday traffic; for months it featured nothing but plugs and portable bathrooms sitting in the dirt.
In cities and suburbs, it’s not uncommon to find charging stations at outlet malls and shopping centers. It makes sense: These places have lots of parking spaces, room for the necessary electrical infrastructure, and stores and restaurants to provide some level of amusement or distraction. If it so happens that you need to go to the REI or Sephora anyway, then so much the better. Mercedes-Benz is trying to class up this setup by putting its luxury charging sites at high-end malls and providing primo, covered parking spaces.
But the game changer is the Waffle House. Businesses have long realized the benefit of adding EV chargers, either as a serendipitous perk for customers who arrive in electric need, or as an enticement for EV owners to patronize their business rather than the competitor with no plugs. Mostly, though, those businesses install Level 2 “destination” chargers that are roughly equivalent to what drivers get in their garage if they pay for the upgrade: 240 volts, or enough to provide 20 to 30 miles of range per hour.
That’s perfect for a hotel, where patrons who snag a charger can wake up the next morning with a full battery, just as they would at home. I made it across sparse Utah country this way. At a grocery store or a restaurant it’s less useful. It’s a pleasant bonus to add a few miles of juice during an errand. What would be better would be filling up the whole battery while you’re inside the Whole Foods.
The problem, however, is timing. Chargers are a shared resource. For optimal EV charging that works for everybody, drivers move their cars as soon as they’re done to open the stall for someone else, which is why many fast-charging operators ding drivers with idle fees if they stay plugged in. So not every activity is a perfect match. It’s pretty annoying to leave your half-filled cart inside Trader Joe’s to go move the car, or to rush through shopping so you finish by the time the battery does. I’ve been through plenty of situations where I couldn’t get back to my Model 3 right away, and so even though it was about to finish charging at 80%, I used the phone app to bump up the limit to 90% or higher to keep the session going.
You know what is a decent match? The Waffle House. You can probably finish your All-Star Special in time, and if you can’t, no problem. This isn’t fine dining; you can leave the table a moment to hop out to the parking lot and unplug the EV.
Putting chargers at the places Americans love to go anyway, whether road tripping or not, would be a wonderful little way to boost their desirability. My native Nebraska has Superchargers co-located with Runzas at towns along the interstate, a welcome trend that must expand. Let Wisconsinites fill the battery while crushing a frozen custard at Culver’s. Give us chargers at the Cracker Barrel so I can finally solve that unholy peg game. Continue the California trend of putting plugs at the In N Out. If the charging stop is someplace you want to go anyway, the minutes required melt away.
Current conditions: The first U.S. heat wave of the year begins today in the West, with a record high of 107 degrees Fahrenheit possible in Redding, California • India is experiencing its earliest monsoon in 16 years• Power was largely restored in southeast Texas by early Wednesday after destructive winds left nearly 200,000 without electricity.
The global average temperature is expected to “remain at or near” the 2-degree Celsius threshold within the next five years, the World Meteorological Organization shared in a new report Wednesday morning. The 2015 Paris Climate Agreement set a warming limit to under 2 degrees C above pre-industrial times, although the WMO’s prediction will not immediately mean the goal has been broken, since that threshold is measured over at least two decades, the Financial Times reports. Still, WMO’s report represents “the first time that scientists’ computer models had flagged the more imminent possibility of a 2C year,” FT writes. Other concerning findings include:
You can find the full report here.
The Federal Emergency Management Agency has been in disarray since its acting administrator was fired in early May for defending the agency before Congress. His successor, David Richardson, began his tenure by threatening staff. According to an internal FEMA memo obtained by The Handbasket, however, the picture is worse than mere dysfunction: Stephanie Dobitsch, the associate administrator for policy and program analysis, wrote to Richardson last week warning him that the agency’s “critical functions” are at “high risk” of failure due to “significant personnel losses in advance of the 2025 Hurricane Season.”
Of particular concern is the staffing at the Mount Weather Emergency Operations Center, which The Handbasket notes contains the nuclear bunker “where congressional leaders were stashed on 9/11,” and which, per Dobitsch, is now “at risk of not being fully mission capable.” FEMA’s primary disaster response office is also on the verge of being unable to “execute response and initial recovery operations and may disrupt life-saving and life-sustaining program delivery,” the memo goes on. Hurricane season begins on Sunday, and wildfires are already burning in the West. You can read the full report at The Handbasket.
The Supreme Court on Tuesday rejected a religious liberty appeal by the San Carlos Apache Tribe to stop the mining company Rio Tinto from proceeding with its plan to build one of the largest copper mines in the world at Oak Flat in Arizona, which the Tribe considers sacred land. Justices Neil Gorsuch and Clarence Thomas said in a dissent that they would have granted the Tribe’s petition, with Gorsuch calling the court’s decision a “grave mistake” that could “reverberate for generations.” The Trump-appointed justice argued that “before allowing the government to destroy the Apaches’ sacred site, this Court should at least have troubled itself to hear their case.”
I traveled to Superior, Arizona, last year to learn more about Rio Tinto’s project, which analysts estimate could extract enough copper to meet a quarter of U.S. demand. “Copper is the most important metal for all technologies we think of as part of the energy transition: battery electric vehicles, grid-scale battery storage, wind turbines, solar panels,” Adam Simon, an Earth and environmental sciences professor at the University of Michigan, told me of the project. But many skeptics say that beyond destroying a culturally and religiously significant site, there is not the smelting capacity in the U.S. for all of Rio Tinto’s raw copper, which the company would likely extract from Oak Flat and send to China for processing. According to court documents, Oak Flat could be transferred to Rio Tinto’s subsidiary Resolution Copper as soon as June 16. In a statement, Wendsler Nosie Sr. of Apache Stronghold — the San Carlos Apache-led religious nonprofit opposing the mine — said, “While this decision is a heavy blow, our struggle is far from over.”
MTA
New York won a court order on Tuesday temporarily preventing the Trump administration from withholding funding for state transportation projects if it doesn’t end congestion pricing, Gothamist reports. The toll, which went into effect in early January, charges most drivers $9 to enter Manhattan below 60th street, and has been successful at reducing traffic and raising millions for subway upgrades. The Trump administration has argued, however, that the toll harms poor and working-class people by “unfairly” charging them to “go to work, see their families, or visit the city.”
The Federal Highway Administration warned New York’s Metropolitan Transportation Authority that it had until May 28 to end the program, or else face cuts to city and state highway funding. Judge Lewis J. Liman blocked the government from the retaliatory withholding with the court order on Tuesday, which extends through June 9, arguing the state would “suffer irreparable harm” without it. Governor Kathy Hochul, a Democrat, celebrated the move, calling it a “massive victory for New York commuters, vindicating our right as a state to make decisions regarding what’s best for our streets.”
European Union countries agreed on Tuesday to dramatically scale back the bloc’s carbon border tariff so that it will cover only 10% of the companies that currently qualify, Reuters reports. The scheme applies a fee on “imported goods that is equivalent to the carbon price already paid by EU-based companies under the bloc’s CO2 emissions policies,” with the intent of protecting Europe-based companies from being undercut by foreign producers in countries that have looser environmental regulations, Reuters writes. The EU justified the decision by noting that the approximately 18,000 companies to which the levy still applies account for more than 99% of the emissions from iron, steel, aluminum, and cement imports, and that loosening the restriction will benefit smaller businesses.
The famous “climate stripes” graphic — which visualizes the annual increases of global average temperature in red and blue bands — has been updated to include oceanic and atmospheric warming. “We’ve had [these] warming estimates for a long time, but having them all in one graphic is what we’ve managed to do here,” the project’s creator, Ed Hawkins, told Fast Company.